This README is based on Mindrocket's "Clueless Coder" standard

Reasons for Fork

• dependencies in requirements.txt 'frozen' to allow repo to age gracefully
• improved README

PREREQUISITES

- Python 3.11.10

- OpenRouter Account

- VSCode & Dev Containers are optional and only of value if you want to tinker with code

If you aren't up and running in 15 minutes, WALK AWAY. OpenRouter Playground is a lot less painful

• If you are comfortable working with venv's, use Option 1
• If you don't see the point of venv's, use Option 2
• Option 3 and Docker. If you know, you know. If you don't know, AVOID

Important Notices

Devcontainer

/.devcontainer/devcontainer.json has deliberately been misnamed ".devcontainer.json". This is to avoid VSCode or Docker launching Devcontainer for a user unfamiliar with the purpose and functionality of Devcontainers. To use, rename the filename to "devcontainer.json"

Defaults

/shared/constants.py contains default OpenRouter settings including default LLMs.

graph TD
    A[Start] --> B[Ensure Git is installed]
    B --> C[Go to working directory]
    C --> D[Clone repo: git clone <repo-url>]
    D --> E{Choose setup option}
    
    E -->|Option 1: Conda/venv| F[Prerequisites:<br>1. Anaconda or Python installed<br>2. Familiarity with command line]
    F --> G[cd into subdirectory]
    G --> H[Create venv:<br>conda create -n myenv python=3.11.10<br>or<br>python -m venv myenv]
    H --> I[Activate venv:<br>conda activate myenv<br>or<br>myenv\Scripts\activate Windows<br>source myenv/bin/activate Unix]
    I --> J[pip install -r requirements.txt]
    J --> K[Ensure Streamlit is installed:<br>pip install streamlit]
    K --> L[streamlit run filename.py]
    
    E -->|Option 2: VSCode + Devcontainer| M[Prerequisites:<br>1. VSCode installed<br>2. Remote-Containers extension in VSCode]
    M --> N[Open repo folder in VSCode]
    N --> O[Ensure .devcontainer folder exists in repo]
    O --> P[Use command palette: Ctrl+Shift+P]
    P --> Q[Select 'Remote-Containers: Reopen in Container']
    Q --> R[Wait for container to build and start]
    
    E -->|Option 3: Docker| S[Prerequisites:<br>1. Docker installed and running<br>2. Familiarity with command line]
    S --> T[cd into subdirectory]
    T --> U[Ensure Dockerfile exists in repo]
    U --> V[docker build -t my-app .]
    V --> W[docker run -p 8501:8501 my-app]
    W --> X[Access app at http&colon;//localhost&frasl;8501]
    
    L --> Y[Access Streamlit app in web browser]
    R --> Z[Follow any additional setup instructions in VSCode terminal]
    X --> AA[End]
    Y --> AA
    Z --> AA

    style A fill:#f9f,stroke:#333,stroke-width:2px,color:#000
    style E fill:#bbf,stroke:#333,stroke-width:2px,color:#000
    style AA fill:#bfb,stroke:#333,stroke-width:2px,color:#000

Loading

🔀 OpenRouter + Streamlit Example App

Starter examples for building LLM apps with Streamlit and OpenRouter, using OpenRouter OAuth PCKE.

Reasons for Fork

• requirements.txt is frozen to allow repo to age gracefully
• improved README

Overview of the App

OpenRouter is a single access point, cost centre and API standard in a landscape of a growing number of LLMs. This repo is a minimum working example of using a single API to access multiple language models, including OpenAI's o1-mini and o1-preview, Anthropic's Claude, Google's Gemini and many more.

Current examples include:

• Chatbot
• File Q&A
• Langchain Quickstart
• Langchain PromptTemplate
• LangChain Search (requires a Serper.dev API key)

Demo App

Running the code

python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
streamlit run Chatbot.py

Setting API keys

Not needed!

• Click the Connect OpenRouter button on the sidebar on the left and your API key will be auto-applied using an OAuth PKCE flow.
• If using LangChain search, enter your Serper API key in the field provided (also on the sidebar on the left of the screen